Security audit
Security checks for vulnerabilities and agentic risk
This cron scheduling skill is mostly purpose-aligned, but it can create persistent scheduled jobs and bulk/delete them with weak documented safeguards.
Review this skill before installing if you rely on scheduled jobs. Use it only with explicit task names or tags, avoid --all unless you have checked the affected jobs, and confirm generated schedules before applying or deleting tasks.
The skill metadata/description mixes unrelated translation/localization use-cases into a cron scheduling skill, indicating copy-paste drift or tampering. This can mislead an agent into invoking the skill for unintended tasks and, because the skill exposes read/exec/write tools, increases the chance of inappropriate command execution under the wrong user intent.
The description states 支持中文交互，无需复杂配置即开即用, while the rest of the file consistently frames supported natural-language patterns as Chinese-only. This is a language/locale constraint presented as a default capability without an explicit user opt-in or documented justification for restricting interaction to a single language.
The skill documents direct deletion of tasks without any confirmation, preview, or recovery guidance. In a scheduling tool, accidental deletion can silently disable recurring operational jobs, reminders, or health checks, creating integrity and availability risks.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
cron-assist template health-check \
--endpoint "https://api.example.com/health" \
--interval 5m --apply
Bulk pause/resume-all commands are presented as normal workflow steps without emphasizing their broad blast radius. An agent following these instructions could suspend all scheduled automations, including critical reminders or monitoring jobs, causing service gaps or missed operations.
The capability-coverage section asserts support for translation/localization scenarios that contradict the rest of the skill. In an agent ecosystem, contradictory capability claims are dangerous because routing and trust decisions may be made from documentation, causing this exec-capable skill to be selected in contexts it was not designed to handle.